Singapore's MAS Publishes Framework to Keep AI Agents Safe in Finance

The Monetary Authority of Singapore has teamed up with major financial institutions and fintech firms to publish a white paper aimed at keeping autonomous artificial-intelligence agents in finance operating within safe boundaries. The paper, titled Safeguards for Agentic Finance at Runtime (SAFR), sets out an industry-built framework designed to let AI agents carry out financial tasks in a way that is safe, secure and dependable. It was produced under BuildFin.ai, a MAS programme that supports the responsible development and rollout of AI tools across the financial sector. The initiative responds to a rapid shift in which AI agents are taking on more autonomous work, at a pace that makes continuous hands-on human oversight impractical. As a result, firms increasingly need real-time controls that keep agent behaviour inside the mandates, policies and risk limits they define, rather than relying solely on after-the-fact checks. By convening banks, insurers and technology providers around a common set of runtime safeguards, MAS aims to encourage innovation while containing risks such as errant transactions, security breaches and unintended actions. The framework reinforces Singapore's positioning as a hub that seeks to lead on both financial technology adoption and its safe governance.